The low phase noise, low supply voltage 1.3 GHz CMOS VCO has been realized by 0.25 µm standard CMOS technology without any trimming and any tuning. The phase noise characteristics of -109 dBc/Hz and -123 dBc/Hz at 100 kHz offset and 500 kHz offset were achieved from carrier, respectively, with 1.3 GHz oscillation frequency at 1.4 V supply voltage. The performance of 1.4 V supply voltage phase noise was superior to that of 2.0 V supply voltage phase noise due to low output impedance current source. The tuning ranges of 13.3%, 16.6%, and 20.1% for 1.4 V, 1.8 V, and 2.0 V supply voltage were achieved, respectively. The amplifier consisted of one pair of PMOS differential stage with large gate length NMOS current source to realize low supply voltage operation and to avoid flicker noise contribution for phase noise. The on-chip spiral inductor consisted of three terminals arranged in a special shape to obtain high Q and small chip area. The power dissipation of this VCO was 22.4 mW without buffer amplifier.

To an extremely difficult problem of finding the maximum likelihood estimates in a specific mixture regression model, a combination of several optimization techniques is found to be useful. These algorithms are the continuation method, Newton-Raphson method, and simplex method. The simplex method searches for an approximate solution in a wider range of the parameter space, then a combination of the continuation method and the Newton-Raphson method finds a more accurate solution. In this paper, this combination method is applied to find the maximum likelihood estimates in a Weibull-power-law type regression model.

This paper describes the scalability of a full-mesh wavelength division multiplexing star-structure network based on an arrayed-waveguide grating router (AWG-STAR). The scalability of the network is examined experimentally. A power penalty of 0.1dB is obtained with a 32-node network and an estimated scalability of up to 100 nodes is confirmed.

A fast-motion-estimation (ME) algorithm called a "breaking-off-search (BOS)" was developed. It can improve processing speed of the full-search (FS) method by a factor of 3.4. The BOS algorithm can not only sometimes achieve better visual quality than FS, but can also solve visual degradation problems associated with conventional fast-ME algorithms whenever picture patterns change (i. e. , presence of scene changes). The power dissipation of a 0.6-µ m CMOS parallel Wallace-tree motion estimator using BOS was reduced to about 281 mW which was 1/28.7 that of the 0.6-µ m CMOS binary-tree motion estimator using FS.

In this letter, we propose a new method to detect faces in color images based on the characterized convex regional relationship. We detect skin and hair likeness regions using the derived skin and hair color models and the convex skin likeness and hair likeness regions are adopted as the characteristic convex regions. Finally, human faces can be detected via their intersection relationship. The proposed algorithm can accomplish face detection in an image including not only single face but also multi-faces and also detect deformed faces efficiently. To validity the effectiveness of the proposed method, we make experiments with various cases.

Since the beginning of the last two decades, many researchers have been involved in the problem of Blind Source Separation (BSS). Whilst hundreds of algorithms have been proposed to solve BSS. These algorithms are well known as Independent Component Analysis (ICA) algorithms. Nowadays, ICA algorithms have been used to deal with various applications and they are using many performance indices. This paper is dedicated to classify the different algorithms according to their applications and performances.

In this letter we propose a new fast matching algorithm that has no degradation of predicted images such as found in the conventional full search (FS) algorithm, so as to reduce the amount of computation of the FS algorithm for motion estimation in real-time video coding applications. That is, our proposing algorithm reduces only unnecessary computations in the process of motion estimation without decreasing the prediction quality compared to the conventional FS algorithm. The computational reduction comes from rapid elimination of impossible motion vectors. In comparison to the FS algorithm, we obtained faster elimination of inappropriate candidate motion vectors using efficient matching units based on image complexity. Experimentally, we demonstrated that the unnecessary computations were removed by about 30% as compared to the other fast FS algorithms.

In this paper, we propose a novel 2D plenoptic function called "acentric panorama view (APV)." This novel 2D plenoptic function samples the panorama scenes without oversampling or undersampling. A single APV can be accelerated by view culling and list-priority rendering algorithm. Multiple APVs with special fields of view, 45, 60, 90, and 120, can be integrated into a larger configuration called augmented APVs, which can augment the walking area in a planar walkthrough environment to form a 4D plenoptic function. In augmented APVs, the acceleration schemes of a single APV can still be applied successfully.

In this paper, we present a new soft-decision iterative decoding algorithm using an efficient minimum distance search (MDS) algorithm. The proposed MDS algorithm is a top-down and recursive MDS algorithm, which finds a most likely codeword among the codewords at the minimum distance of the code from a given codeword. A search is made in each divided section by a "call by need" from the upper section. As a consequence, the search space and computational complexity are reduced significantly. The simulation results show that the proposed decoding algorithm achieves near error performance to the maximum likelihood decoding for any RM code of length 128 and suboptimal for the (256, 37), (256, 93) and (256, 163) RM codes.

Fully integrated VCO using the "turbo-charger" technique to improve phase noise characteristics is presented. The phase noise degradation of relatively lower oscillation frequency in tuning range was caused by oscillation amplitude lowering due to large total capacitance. On the other hand, the phase noise degradation of relatively higher frequency in tuning range was caused by excess current noise. A new "turbo-charger" circuit increased operation current to obtain sufficient transconductance of amplifier when oscillation frequency was lower to improve phase noise characteristics. The phase noise of VCO employing this technique was extremely low and stable, below -140-dBc/Hz at 3-MHz offset from oscillation frequency, in wide oscillation frequency range, approximately 200-MHz for 1200-MHz oscillation. This VCO was operated with 5.8-7.4-mA current consumption at 3-V supply voltage. The manufacturing process was 0.6-µm SiGe BiCMOS.

A 6 GHz integrated VCO using SiGe BiCMOS process has been studied. The integrated inductors were realized by third metal with 3 µm thickness aluminum and its Q=20 at 6 GHz. The amplifier consisted of bipolar transistor. Tuning range was 38% with 0 V to 3 V tuning voltage. Phase noise of -100 dBc/Hz was obtained at 1 MHz offset from carrier frequency. The current consumption of VCO was 4.9 mA at 3 V power supply.

An English auction is the most familiar type of auctions. Generally, an electronic auction has mainly two entities, the registration manager (RM) who treats the registration of bidders, and the auction manager (AM) who holds auctions. Before starting an auction, a bidder who wants to participate in English auction is registered to RM with her/his information. An electronic English auction protocol should satisfy the following nine properties, (a) Anonymity, (b) Traceability, (c) No framing, (d) Unforgeability, (e) Fairness, (f) Verifiability, (g) Unlinkability among plural auctions, (h) Linkability in an auction, and (i) Efficiency of bidding. Furthermore from the practical point of view we add two properties (j) Easy revocation and (k) One-time registration. A group signature is adapted to an English auction in order to satisfy (a), (b), and (f). However such a direct adoption suffers from the most critical drawback of efficiency in group signatures. In this paper we propose more realistic electronic English auction scheme, which satisfies all of these properties without using a group signature. Notable features of our scheme are: (1) both of bidding and verification of bids are done quite efficiently by introducing a bulletin board, (2) both properties (j) Easy revocation and (k) One-time registration are satisfied.

We propose a new and fast full search (FS) motion estimation algorithm for video coding. The computational reduction comes from sequential rejection of impossible candidates with derived formula and subblock norms. Our algorithm reduces more the computations than the recent fast full search (FS) motion estimation algorithms.

A high isolation hybrid circuit composed of a pair of transformers, a voltage control resistance (VCR) circuit and an automatic impedance control device is designed for a two-wire full duplex cable modem to adapt variable line impedance. A binary frequency shift keying (BFSK) cable modem using the new hybrid circuit with an isolation of 52 dB to 58 dB in the line impedance variation range of 400 to 950 ohm is demonstrated. The isolation of the new hybrid circuit is increased by more than 30 dB over the traditional hybrid circuit for a two-wire full duplex modem in the preset line impedance range.

This paper deals with a method of registration and superimposition of a coronary arterial tree on a myocardial SPECT (Single Photon Emission Computed Tomography) image. We can grasp the myocardial function more easily in connection with the shape of the coronary arterial tree. The superimposed image is easily obtainable through some manual pointing on coronary angiograms (CAG) followed by an automatic matching method: First, a rough shape model of left ventricle is estimated by using SPECT data. This 3-D left ventricular model is projected on a pair of bi-plane CAG images. We can obtain two 2-D coronary images on bull's eye map by scanning the left ventricular surface projected on CAG. By maximizing a matching degree between two 2-D coronary images, registration between CAG and SPECT is performed. Finally the superimposed image is obtained by integrating two 2-D coronary images and bull's eye image of SPECT. We validated our method by numerical experiments with artificial data set and also applied it to two clinical data sets.

As LSIs are two-dimensional structures, the number of external pins increases at a lower rate than the corresponding increase in the number of gates on the LSI. Therefore, the number of flip-flops on a scan path increases as the density of gates on LSIs rises, resulting in longer test application times. In this paper, three novel DFT strategies aimed at reducing test application time are proposed. DFT strategy 1 is a full scan design method with test point insertion, DFT strategy 2 is a partial scan design method, and DFT strategy 3 is a partial scan design method with test point insertion. Experimental results show that these DFT strategies reduced the test application times by 45% to 82% compared with conventional full scan design methods.

Autonomous Information Service System is a proposition made to cope with the continuously changing conditions of service provision and utilization in current information systems. The faded information field (FIF), sustained by push/pull mobile agent technology, is such a distributed architecture that brings high-assurance of the system through a balanced selective replication of the information. When the demand changes, the information environment is restructured so that the same response time to services for all unspecified users can be achieved whatever the demand volume. However, once the structure is fixed, dispatching the randomly incoming requests on the FIF is still required to guarantee the same quality of service. Our goal is to warrant the autonomous dispatching of the pull mobile agents to adjust the continuously evolving arrival distribution of the demand to the current information environment. In this paper, we explain the concepts and realization of autonomous navigation under the goal of an autonomous load balancing of the pull mobile agent volume in the FIF structure. The appropriateness of this method for FIF environments has been shown by simulation.

Beam reconfiguration by structural reconfigurable antenna, such as the small multi-panel reconfigurable reflector antenna, has an aspect of great concern, that is the effects due to the use of a number of small panels to form the reflecting surface. It is thus a matter of great interest to numerically investigate all possible factors affecting the performance of this type of antenna such as: neighboring panels blocking, diffraction. The "null-field hypothesis" and PTD are employed to account for the effects of both phenomena on the main beam steering ability and the cross-polar level. In addition, the transformation of the polygonal flat domains into the square domains is applied in calculating the PO radiation field due to the various irregular polygonal flat sections of the arbitrary initial approximate reflector e.g., the flat circular reflector and the paraboloidal reflector. It is found that the main contribution to the total cross polarization is depolarization due to the finite size of the panels. The maximum cross-polar gain predicted using PTD is around -30 dB. The blocking effect has minor influence on cross-polarization. Both effects cause distortion on the co-polar pattern for the observer far from boresight but blocking has more influence than edge diffraction. Both effects have minor influence on the co-polar gain. The co-polar gain has variation of less than or equal to 0.07 dB in the flat case and 0.16 dB in the paraboloid case.

In this paper, we present deterministic parallel algorithms for the convex hull of sorted points and their application to a related problem. The algorithms are proposed for the coarse grained multicomputer (CGM) model. We first propose a cost optimal parallel algorithm for computing the problem with a constant number of communication rounds for n/p p2, where n is the size of an input and p is the number of processors. Next we propose a cost optimal algorithm, which is more complicated, for n/p pε, where 0 < ε < 2. From the above two results, we can compute the convex hull of sorted points with O(n/p) computation time and a constant number of communication rounds for n/p pε, where ε > 0. Finally we show an application of our convex hull algorithms. We solve the convex layers for d lines in O((n log n)/p) computation time with a constant number of communication rounds. The algorithm is also cost optimal for the problem.

Multi-level divide-and-conquer (MDC) is a generalized divide-and-conquer technique, which consists of more than one division step organized hierarchically. In this paper, we investigate the paradigm of the MDC and show that it is an efficient technique for designing parallel algorithms. The following parallel algorithms are used for studying the MDC: finding the convex hull of discs, finding the upper envelope of line segments, finding the farthest neighbors of a convex polygon and finding all the row maxima of a totally monotone matrix. The third and the fourth algorithms are newly presented. Our discussion is based on the EREW PRAM, but the methods discussed here can be applied to any parallel computation models.